Hi all ,

Somehow I want to try (I think) an EF80 as a CF , but
it seems there's not much on the net about pentode CF's .

What should be the best CF with an EF80 ??

1) Triode connected (should at least work I think) .

2) Pentode connected , G2 decoupled to ground .

3) Pentode connected G2 decoupled to cathode .

Somehow my guts tell me to go for 3) , but some
theory won't harm I think .....

T.I.A.

Ronald .

Ronald wrote:

Hi all ,

Somehow I want to try (I think) an EF80 as a CF , but
it seems there's not much on the net about pentode CF's .

What should be the best CF with an EF80 ??

1) Triode connected (should at least work I think) .

EF80 has a µ = 50 in triode and makes an excellent CF.




2) Pentode connected , G2 decoupled to ground .

But for pentode you must have G2 bypassed with an electro to k.
The G2 supply resistor then loads the cathode circuit.
But it works.



3) Pentode connected G2 decoupled to cathode .

Well that'd be the same as 2)



Somehow my guts tell me to go for 3) , but some
theory won't harm I think .....

I'd use triode, 1), Ro will be 1/gm = 1/0.007 = 140 ohms,
thd at a volt will be hard to measure.

Patrick Turner.


T.I.A.

Ronald .

Thanks Patrick ,

I think I'll try triode (first) then .....
It's the most simple solution anyway ;-))
Maybe a little R (B+ to G2) to prevent for oscilation .....

Ronald .

"Patrick Turner" schreef in bericht
...


Ronald wrote:

Hi all ,

Somehow I want to try (I think) an EF80 as a CF , but
it seems there's not much on the net about pentode CF's .

What should be the best CF with an EF80 ??

1) Triode connected (should at least work I think) .

EF80 has a µ = 50 in triode and makes an excellent CF.




2) Pentode connected , G2 decoupled to ground .

But for pentode you must have G2 bypassed with an electro to k.
The G2 supply resistor then loads the cathode circuit.
But it works.



3) Pentode connected G2 decoupled to cathode .

Well that'd be the same as 2)



Somehow my guts tell me to go for 3) , but some
theory won't harm I think .....

I'd use triode, 1), Ro will be 1/gm = 1/0.007 = 140 ohms,
thd at a volt will be hard to measure.

Patrick Turner.


T.I.A.

Ronald .

"Patrick Turner" wrote

1) Triode connected (should at least work I think) .

EF80 has a µ = 50 in triode and makes an excellent CF.

2) Pentode connected , G2 decoupled to ground .

But for pentode you must have G2 bypassed with an electro to k.
The G2 supply resistor then loads the cathode circuit.
But it works.

3) Pentode connected G2 decoupled to cathode .

Well that'd be the same as 2)

er...why? 2) is the same as 1), actually.

Ronald, for a CF, there is no signal at the anode, so bypassing to
ground is the same thing as connecting to the anode, except you
introduce the complication of a frequency-dependent connection if
you use a cap to ground instead of the simpler triode connection.

Discussion of a pentode CF would be interesting, so why not try 1)
and 3) and report back?

The interesting aspect is that you would be replacing the
degeneration of a triode with extra cathode feedback applied to the
pentode, due to its higher open-loop gain. In some respects you
might expect the net result to be similar.

So what would be the difference in output impedance, and distortion
spectrum, and maximum voltage swing? Is the ideal operating point
and load different? Plenty to wonder about.

cheers, Ian

Ronald wrote:

Thanks Patrick ,

I think I'll try triode (first) then .....
It's the most simple solution anyway ;-))
Maybe a little R (B+ to G2) to prevent for oscilation .....

Yup, maybe 270 ohms.
Also, g3 can be taken to k or a.

Make sure you have the anode supply well bypassed to 0V
from near the anode connection lest the inductance of a long lead
make it oscillate at some RF.

The EF80 in triode is similar to a paralleled 12AT7.

You can use another EF80 for a high impedance current sink
instead of a cathode resistor.
Say you have a 250v supply, with the drive to the CF from a fixed bias
voltage at
say +130v.

Just set up the CS with unbypassed Rk of a few hundred ohms to set the
current,
take g1 and g3 to 0V, take G2 to +130v, and let the anode connect to the

cathode of the CF. The CS will be a dynamic resistance of about
1Meg+ at say 5 mA.
Then the thd of the CF will definately be hard to measure at a volt.
The cap coupled load then becomes the only substantial load the
CF has to power.

A bjt can also make a nice CS for CF.

Even a choke is good, and then the cathode voltage is down near 0V,
since the choke has low dcr.
The grid may then be biased at 0V, like any other tube,
and voltage biasing of the heater supply isn't needed.

Another way is to take the CS to a -ve supply voltage...
Lot of ways to do it.

Patrick Turner.




Ronald .

"Patrick Turner" schreef in bericht
...


Ronald wrote:

Hi all ,

Somehow I want to try (I think) an EF80 as a CF , but
it seems there's not much on the net about pentode CF's .

What should be the best CF with an EF80 ??

1) Triode connected (should at least work I think) .

EF80 has a µ = 50 in triode and makes an excellent CF.




2) Pentode connected , G2 decoupled to ground .

But for pentode you must have G2 bypassed with an electro to k.
The G2 supply resistor then loads the cathode circuit.
But it works.



3) Pentode connected G2 decoupled to cathode .

Well that'd be the same as 2)



Somehow my guts tell me to go for 3) , but some
theory won't harm I think .....

I'd use triode, 1), Ro will be 1/gm = 1/0.007 = 140 ohms,
thd at a volt will be hard to measure.

Patrick Turner.


T.I.A.

Ronald .

Ian Iveson wrote:

"Patrick Turner" wrote

1) Triode connected (should at least work I think) .

EF80 has a µ = 50 in triode and makes an excellent CF.

2) Pentode connected , G2 decoupled to ground .

But for pentode you must have G2 bypassed with an electro to k.
The G2 supply resistor then loads the cathode circuit.
But it works.

3) Pentode connected G2 decoupled to cathode .

Well that'd be the same as 2)

er...why? 2) is the same as 1), actually.

well, the same as I said, not what Ron said.



Ronald, for a CF, there is no signal at the anode, so bypassing to
ground is the same thing as connecting to the anode, except you
introduce the complication of a frequency-dependent connection if
you use a cap to ground instead of the simpler triode connection.

But the screen needs to be at a much higher potential than the cathode
lest the tube be cut off, but yes, if the potential is the B+ as for the
anode, then its triode
connected.

But for eal pentode action, g2 could be 50v below Ea, bypassed to k and
with say 50k supply R
from B+. Hence the cathode sees the 50k as a load, in addition
to its Rk to 0V or a -ve supply.
Rk would be about 20k for 5 mA with 100v across Rk.
RL total is therefore 50k plus 20k plus the outside world RL of say 50k,

so total in parallel = 11k, and open loop gain =
Gm x RL = 0.007 x 11,000 = 77,
and not much above what triode operation would offer.
So the closed loop gain reduction to just under 1.0
conveys few benefits just because you have used pentode operation.






Discussion of a pentode CF would be interesting, so why not try 1)
and 3) and report back?

The interesting aspect is that you would be replacing the
degeneration of a triode with extra cathode feedback applied to the
pentode, due to its higher open-loop gain. In some respects you
might expect the net result to be similar.

The pentode connection is like an application of positive feedback
to what would otherwise be a low µ triode.
The pentode µ of EF80 = gm x Ra, and
at 5 mA, µ = approximately 0.007 x 500,000 = 3,500.
But yes there is NFB around the g2 circuit.
But as IRL causes a rise in VRL, if g1 goes +ve, the
g2 is bootstrapped off k, and lifts the g2 voltage, thus tending to
keep the effect of a reducing a to k voltage from reducing Ia.

I might add that one can have a CF whose anode is supplied from another
CF
connected above the main signal CF, and the purpose of the top CF is to
raise the anode signal of the main signal CF. Its another
bit of bootstrapping.

See Allen Wright's site at http://www.vacuumstate.com
and look for "super cathode follower"



So what would be the difference in output impedance, and distortion
spectrum, and maximum voltage swing? Is the ideal operating point
and load different? Plenty to wonder about.

I'm using 12AU7 CF output buffers on my preamp.
I have MJE350 CCS, about 20Meg actual R, high enough.
thd is less than 0.01% at a volt, more than I ever use.

It is simple and sounds well, but there are other ways
if we have a pentode.
If the g2 supply is CCS with good bypassing, and cathode current is CCS,
then the
pentode open loop gain will be about 3,000.
If the outside world load is 50k, OLG
= gm x RL = 0.007 x 50,000 = 350,
and the CF connection reduces this gain to less than 1.0,
so if you had 0.5% thd at 3v with 50k load with
common cathode, or OLG, there'd be (0.5/350) % thd with CF, or 0.0014%.

Triode would give you OLG about 0.25% thd at 3v output, ( based on
about
5% at 60v out, with a nearly linear reduction of thd with vo )
I am assuming triode thd = 1/2 the pentode thd with 50k loading.
So with a gain reduction of say 45 the thd in triode CF at 3v would be
around
0.056%, which is 4 times more than the pentode performance
with all the CCS bells and whistles attatched to make it work with
maximum FB due to maximum
OLG.
Without such CCS bells and whistles the pentode performance would sag
back to triode.
But hey, 0.056% thd at 3vo isn't a bad figure.
if the application is a preamp and all we want is 0.3v, then thd
will be 0.0056%; who cares, we are well over the line for
technical "good enoughness", and the sound the EF80
gives will be blameless.

Anyway, I make rough estimations here.
Nothing beats building and listening and measuring.

Patrick Turner.







cheers, Ian

Hi Patrick ,

3) Pentode connected G2 decoupled to cathode .

Well that'd be the same as 2)

And what if I make R (B+ to G2) somwhat bigger , lets say 27k ??
(assuming Ig2 = 1 to 1.5 mA)
Roughly the output signal is at G2 too then ......

Make sure you have the anode supply well bypassed to 0V

I figured that out , but isn't the G2 supply even more important ??

from near the anode connection lest the inductance of a long lead
make it oscillate at some RF.

That's what I always do , the last PSU cap close to the stage and
leads as short as possible ofcourse ;-)

The EF80 in triode is similar to a paralleled 12AT7.

That's nice to know , I couldn't find trioded data on this tube .

You can use another EF80 for a high impedance current sink
instead of a cathode resistor.

Maybe I could , but I don't have much room ......
The CF is for a (mono line level) sub-woofer output .
The plan is to build a stereo ECL86 SE amp into a PC .
I expect about 3W from that , so sub-bass is gonna be hard I think ,
so I planned an active sub-woofer (single woofer with dual coil) in
a transmission line cabinet , tuned at about 45Hz .
That should make a nice multi-media system I think ;-) , even if the
sub gets a simple "gain-clown" type of amp .

A bjt can also make a nice CS for CF.

I have some IRF620 FET's IIRC . Maybe that's an option .....

Even a choke is good, and then the cathode voltage is down near 0V,
since the choke has low dcr.
The grid may then be biased at 0V, like any other tube,
and voltage biasing of the heater supply isn't needed.

Sure , but a choke inside a PC .... I don't know ......
And a desent choke might be to big anyway .

Lot of ways to do it.

Sure , but I don't have room for extra tubes .
To get an idea you could take a look at :
http://www.angelfire.com/electronic2...s/casemod.html
The EF80 should replace the EC92 in the schematic at the end of
the page . The EC92 (= 1/2 12AT7) would work too , but using
a 9-pin socket suits me better and an EF80 has a shield so placing
it close to the power tranny might be less a problem ..... (?)

Ronald .

Ronald wrote:

Hi Patrick ,

3) Pentode connected G2 decoupled to cathode .

Well that'd be the same as 2)

And what if I make R (B+ to G2) somwhat bigger , lets say 27k ??
(assuming Ig2 = 1 to 1.5 mA)
Roughly the output signal is at G2 too then ......

It is the same signal if the bypass cap is say 220 uF.



Make sure you have the anode supply well bypassed to 0V

I figured that out , but isn't the G2 supply even more important ??

Its class A. The DC voltage between g2 and k will stay put with signal
applied
for small music signals.
Place a voltmeter across the bypass cap with signal.
The dcv will change only slightly.




from near the anode connection lest the inductance of a long lead
make it oscillate at some RF.

That's what I always do , the last PSU cap close to the stage and
leads as short as possible ofcourse ;-)

I have a couple of hundred 2 uF polyester caps good for 500v.
They do the trick in addition to ther electros.



The EF80 in triode is similar to a paralleled 12AT7.

That's nice to know , I couldn't find trioded data on this tube .

Try Duncans amp pages, go tube data, there are a few links.....

I found and saved a pdf 2 nights ago when looking for
E184F/6EJ7 pentode data.
I also found EF80/6BX6.





You can use another EF80 for a high impedance current sink
instead of a cathode resistor.

Maybe I could , but I don't have much room ......
The CF is for a (mono line level) sub-woofer output .
The plan is to build a stereo ECL86 SE amp into a PC .
I expect about 3W from that , so sub-bass is gonna be hard I think ,
so I planned an active sub-woofer (single woofer with dual coil) in
a transmission line cabinet , tuned at about 45Hz .
That should make a nice multi-media system I think ;-) , even if the
sub gets a simple "gain-clown" type of amp .

A bjt can also make a nice CS for CF.

I have some IRF620 FET's IIRC . Maybe that's an option .....

I've tried HV fets for CCS, but they are delicate, one little fault, and
poof,
they are gone.
I like to use MJE340 and MJE350 for signal duties.
If you have 2.2k for Re, and say 5mA, then that's 11v,
then the base emitter drop of 0.65v, so if you can keep
the voltage between the base input and end of the Re at a constant
11.65 v, then the resistance looking into the
collector circuit is enormous, may megohms.
One transistor is all i need.





Even a choke is good, and then the cathode voltage is down near 0V,
since the choke has low dcr.
The grid may then be biased at 0V, like any other tube,
and voltage biasing of the heater supply isn't needed.

Sure , but a choke inside a PC .... I don't know ......
And a desent choke might be to big anyway .

Inside a PC? tubes? OK, foget chokes, that single bjt is looking good.



Lot of ways to do it.

Sure , but I don't have room for extra tubes .
To get an idea you could take a look at :
http://www.angelfire.com/electronic2...s/casemod.html
The EF80 should replace the EC92 in the schematic at the end of
the page . The EC92 (= 1/2 12AT7) would work too , but using
a 9-pin socket suits me better and an EF80 has a shield so placing
it close to the power tranny might be less a problem ..... (?)

Ronald .

In your schematic below the PC box
with two big round eyes and a horrible mouth,
you have two PCL86 triodes rigged as CF to each accept the
signals from each stereo channel.

But the CF cathodes are tied together, and this this won't work very
well because the output resistance is so low of each CF that they will
fight
each other; one might go +ve, the other -ve, and the tubes will overload,

or at least distort the signals.

What you should have is well separated cathodes each with an Rk
of say 100k to 0V, and then have 22k from each cathode coming to a join
point after which
you can rig the second order feedback filter where you have the EC92.
A single 12AT7 CF for the mono output would be quite adequate.
Or a trioded EF80 would be all I'd use.

I recently built a sub for a client with the filter box and used
a pair of darlington pair connected emitter follower input buffers,
then 2simple RC filter stages, another buffer and so on.
I found The passive RC gave a more rounded knee of treble attenuation,
but I had 5 RC filters all up and by 200 Hz the slope was over 24
dB/octave,
but at -3dB the phase shift wasn't as severe as with maximally flat sharp

second order filters.
You may need to use the second CF triode for a second CF filter.
With only the sub connected and the crossover set at 50Hz
it should be impossible to hear intelligible speach or singing.
It should be just a rumble, sounds weird, I know, but
with the rest of the music it adds ambience which was in the recording.

Like all subs, its position is critical, since bass frequency response in
rooms is very
patchy, and prone to many unwanted phase effects.
I set my clients sub up with the aid of a pink noise source and mic where
he sits,
but it still wasn't easy to get right.
Luckily there was a position where it wasn't a room obstacle.

For various reasons many ppl don't like subs, settling for
reasonable bass from small floorstanders, rather than fake bass
from a badly done sub.
I have two bins 135L with old huge Electrovoice 12" drivers.
These match the boxes ok and give the best bass south of the equator
imho.

but for the latest sub project for a client I used a Peerless 12" XLS in
86 litres tuned to around 30 Hz
with a port of 400 mm. It was much better than a previous sub I did with
130 litres, and same driver.

Patrick Turner.

Hi Patrick ,

Try Duncans amp pages, go tube data, there are a few links.....

I found and saved a pdf 2 nights ago when looking for
E184F/6EJ7 pentode data.
I also found EF80/6BX6.

I did , but didn't find exact data on triode connection .
No big deal for a CF I think ;-) And if I realy need the curves or
something I could measure them like I did for the 1G6 , nice tube btw !!


I have some IRF620 FET's IIRC . Maybe that's an option .....

I've tried HV fets for CCS, but they are delicate, one little fault, and
poof,
they are gone.
I like to use MJE340 and MJE350 for signal duties.

Ok , but it's "just" multi-media so I like to keep the budget a little tight
..
It's more like just good looking to have some tube behind that "window" .
So I think I'll keep it simple at first , just the 27k (I have) .
It'll give about 4.5mA .

Inside a PC? tubes? OK, foget chokes, that single bjt is looking good.

I might even need to shield the OPT's I think , no problem to do so .....

In your schematic below the PC box
with two big round eyes and a horrible mouth,
you have two PCL86 triodes rigged as CF to each accept the
signals from each stereo channel.

I thought to make a sort of mixer this way (L+R) and since there isn't
much stereo in the bass .......

But the CF cathodes are tied together, and this this won't work very
well because the output resistance is so low of each CF that they will
fight
each other; one might go +ve, the other -ve, and the tubes will overload,

or at least distort the signals.

Now you say so it sounds logical .....

What you should have is well separated cathodes each with an Rk
of say 100k to 0V, and then have 22k from each cathode coming to a join
point after which
you can rig the second order feedback filter where you have the EC92.

Hmm , makes sence !! But paralleling both triodes and making a op-amp
type of mixer might be possible too ??
I saw something like that at the Tube Cad Journal , mixing L+R and use
the same R from the anode back to the grid .
http://www.tubecad.com/april_may2001/page22.html
Normaly C-Miller might be a problem that way , but who cares about that
for a sub-woofer output ......
And since the phase tunrs in such a mixer stage I think it's better to take
the L and R signal from the cathodes of the 5687 .
Loosing the gain is more like a pro than a con !!

A single 12AT7 CF for the mono output would be quite adequate.
Or a trioded EF80 would be all I'd use.

I have some nice looking , good testing (leak and emmision) EF80 so
I'll go for that tube I think .

You may need to use the second CF triode for a second CF filter.

In the "power"-amp there's only a first order (Butterworth) filter that'll
match the 2nd order in the sub-filter quite well I think as long as I tune
the speakers connected to the "power"-amp at the same frequency in
a clossed box with a Qtc of 0.707 . That's also a first order butterworth
filter , isn't it ?? And if you connect 2 the same 1st order Butterworth
filters in series you'll get a Linkwitz-Riley slope . That's why I made the
sub-filter a Linkwitz-Riley filter .

With only the sub connected and the crossover set at 50Hz
it should be impossible to hear intelligible speach or singing.
It should be just a rumble, sounds weird, I know, but
with the rest of the music it adds ambience which was in the recording.

I know , I made a passive sub /satalite system in the past , Fc at 125Hz.

For various reasons many ppl don't like subs, settling for
reasonable bass from small floorstanders, rather than fake bass
from a badly done sub.

A good sub isn't a problem , but the common trash you'll see now days .....
Lost of ppl don't even take the time to realize what they're listening to or
if it even sound like the artist ment to .......
Sub or not , if there's no bass in a recording I don't want to hear bass .
And if there is bass it should be in the right proportion , not like those
youngsters with their "boom-boom" cars !!
If they run out of gaz they'll kick the back window and drive home on
the power of the sub (giving me a headake !!)

Ronald .

Hi Ian ,

Thanks for the reply !!
Most points I discussed with Patrick already I think , but .....

Ronald, for a CF, there is no signal at the anode, so bypassing to
ground is the same thing as connecting to the anode, except you
introduce the complication of a frequency-dependent connection if
you use a cap to ground instead of the simpler triode connection.

Yes , I see now that I posted b4 thinking ..... ;-))
The only thing is U-g2 (dc) might be better filtered / regulated , less
ripple I mean .

Discussion of a pentode CF would be interesting, so why not try 1)
and 3) and report back?

Maybe I will try 1) and 3) , but as you can read in my other posts
it's just for a sub-woofer output , so the 1 giving the best bass is
good for me , but might not be the over all best sounding .....

Basicly the idea of using a pentode as CF came from the "Kimmel-
type" mu-stage . Kimmel uses a pentode connected pentode on
top of the triode , driven like a mu-follower .
Somehow I like that stage , but the EF80 isn't good for that !!
An EF184 is way better there ......

Best regards ,

Ronald .

"Ronald" wrote

The only thing is U-g2 (dc) might be better filtered / regulated ,
less
ripple I mean .

Yes..."power supply rejection" is the common term. I thought so too
after I posted. Worth trying that too then!

Basicly the idea of using a pentode as CF came from the "Kimmel-
type" mu-stage . Kimmel uses a pentode connected pentode on
top of the triode , driven like a mu-follower .
Somehow I like that stage , but the EF80 isn't good for that !!
An EF184 is way better there ......

Right. It's a good paper, and inspired me to try a mosfet
mu-follower. But complicated circuits can have complicated
limitations.

cheers, Ian

Ronald wrote:

Hi Patrick ,

Try Duncans amp pages, go tube data, there are a few links.....

I found and saved a pdf 2 nights ago when looking for
E184F/6EJ7 pentode data.
I also found EF80/6BX6.

I did , but didn't find exact data on triode connection .

OK, nor could I and some of the links from Duncan's pages don't work.
The internal amplificayion factor from g1 g2 is given as 50.
I measured 6BX6 with triode µ =50, Ra around 12k.




No big deal for a CF I think ;-) And if I realy need the curves or
something I could measure them like I did for the 1G6 , nice tube btw !!

I have some IRF620 FET's IIRC . Maybe that's an option .....

I've tried HV fets for CCS, but they are delicate, one little fault, and
poof,
they are gone.
I like to use MJE340 and MJE350 for signal duties.

Ok , but it's "just" multi-media so I like to keep the budget a little tight

The bjts are TO26 types and about $2 each from most outlets.
Its a very cheap and effective solution.



.
It's more like just good looking to have some tube behind that "window" .
So I think I'll keep it simple at first , just the 27k (I have) .
It'll give about 4.5mA .

Inside a PC? tubes? OK, foget chokes, that single bjt is looking good.

I might even need to shield the OPT's I think , no problem to do so .....

In your schematic below the PC box
with two big round eyes and a horrible mouth,
you have two PCL86 triodes rigged as CF to each accept the
signals from each stereo channel.

I thought to make a sort of mixer this way (L+R) and since there isn't
much stereo in the bass .......

That's right, you want to extract the bass signal from the combined
L+R stereo signal.



But the CF cathodes are tied together, and this this won't work very
well because the output resistance is so low of each CF that they will
fight
each other; one might go +ve, the other -ve, and the tubes will overload,

or at least distort the signals.

Now you say so it sounds logical .....

What you should have is well separated cathodes each with an Rk
of say 100k to 0V, and then have 22k from each cathode coming to a join
point after which
you can rig the second order feedback filter where you have the EC92.

Hmm , makes sence !! But paralleling both triodes and making a op-amp
type of mixer might be possible too ??

Yes, its not quite as bad, since the output would be from the common anodes.
This has been done, but the right way is to have two independant CF,
with the two joined resistors, and the following second order filter.

I saw something like that at the Tube Cad Journal , mixing L+R and use
the same R from the anode back to the grid .
http://www.tubecad.com/april_may2001/page22.html
Normaly C-Miller might be a problem that way , but who cares about that
for a sub-woofer output ......

The tube cad method is ok because the LR signals are summed outside the
triode, and applied to the grid as one signal.
Then if the shunt feedback version is used, the
gain of the triode can be tailored to suit the rest of the circuit.

But remember such a stage inverts the phase, and perhaps you should try to have
phase coherence between bass and sub signals.



And since the phase tunrs in such a mixer stage I think it's better to take
the L and R signal from the cathodes of the 5687 .
Loosing the gain is more like a pro than a con !!

You may need some gain since you will need to have a separate bass level control

to get the level of sub-bass about right with the
rest of the sound.

The signal could be summed via 100k resistors from each anode
of the 5687 and the junction of these R can have a cap to ground with a pole at
about
200Hz.
The R junction is also taken to a single CF to make a buffer before the second
order filter
with pole at say 150, and you will have 18 dB/octave slope above 300 Hz.

Another RC passive filter after the second CF will give you 24 dB/octave.

The only thing missing is a gain control, but the level of bass recoverd
as suggested will be higher that fed to your PL86.

The sub power amp can have its sensitivity set up
to be able to have a gain control in front of it,
and after the filter to remove treble.



A single 12AT7 CF for the mono output would be quite adequate.
Or a trioded EF80 would be all I'd use.

I have some nice looking , good testing (leak and emmision) EF80 so
I'll go for that tube I think .

You may need to use the second CF triode for a second CF filter.

In the "power"-amp there's only a first order (Butterworth) filter that'll
match the 2nd order in the sub-filter quite well I think as long as I tune
the speakers connected to the "power"-amp at the same frequency in
a clossed box with a Qtc of 0.707 . That's also a first order butterworth
filter , isn't it ?? And if you connect 2 the same 1st order Butterworth
filters in series you'll get a Linkwitz-Riley slope . That's why I made the
sub-filter a Linkwitz-Riley filter .

I get confused with filter descriptions.

But for a passive RC type with fairly rapid roll off with ultimate
attenuation of 12 dB/octave, you can have a low impedance source
such as a cathode follower driving 10k with 0.068 uF to 0V,
giving a pole at 234Hz.
Then a following RC section with 47k and 0.015 uF for a pole at 212 Hz.
The second RC filter has 5 times the impedance value of the first,
so they act to give a reasonably sharper roll off than if you had two RC filters

with the same values.
The combined -3dB pole will be at around 150 Hz, and
by 500 Hz the slope has become steep.
Its naturally damped, and simple.
So you can have CF buffer, two RC filters, another CF buffer, two more RC
sections,
and another CF buffer, and a gain pot after that and you haven't used any loop
NFB,
the distortion will be negligible, an in any case the higher products are
attenuated by the filters.



With only the sub connected and the crossover set at 50Hz
it should be impossible to hear intelligible speach or singing.
It should be just a rumble, sounds weird, I know, but
with the rest of the music it adds ambience which was in the recording.

I know , I made a passive sub /satalite system in the past , Fc at 125Hz.

For various reasons many ppl don't like subs, settling for
reasonable bass from small floorstanders, rather than fake bass
from a badly done sub.

A good sub isn't a problem , but the common trash you'll see now days .....
Lost of ppl don't even take the time to realize what they're listening to or
if it even sound like the artist ment to .......
Sub or not , if there's no bass in a recording I don't want to hear bass .
And if there is bass it should be in the right proportion , not like those
youngsters with their "boom-boom" cars !!
If they run out of gaz they'll kick the back window and drive home on
the power of the sub (giving me a headake !!)

Yes, but research into sub-woofer powered sports cars with solar roof panels
is underway.

They are training the bass player and drummer to provide the right
type of sound to use, and design of suitable ear muffs for the driver
is just about done.
One British horn speaker maker was contracted to provide a horn powered
protoype,
but they had trouble getting a permit from the transport authorities,
since it couldn't fit under the bridges.
Then someone said it'd be more effective to use sub powered barges, and limit
them to use in countryside canals, but the Society Against Hearing Damage in
Tortoises
gained an injunction in the High Court against underwater subs
and underwater horns.

Progress continues in the face of obstacles.

Patrick Turner.




Ronald .

Hi Pat ,

OK, nor could I and some of the links from Duncan's pages don't work.

Most of the time I go directly to Frank's .....
http://www.mif.pg.gda.pl/homepages/frank/
There are many tube not linked to Duncan's !!

The internal amplificayion factor from g1 g2 is given as 50.
I measured 6BX6 with triode µ =50, Ra around 12k.

That would give a S = 4.16 mA/V , indeed pretty close to a 12AT7 and
at least good enough for a CF .

The bjts are TO26 types and about $2 each from most outlets.
Its a very cheap and effective solution.

Ok , $2 on a entire project is peanuts .....

I saw something like that at the Tube Cad Journal , mixing L+R and use
the same R from the anode back to the grid .
http://www.tubecad.com/april_may2001/page22.html

The tube cad method is ok because the LR signals are summed outside the
triode, and applied to the grid as one signal.
Then if the shunt feedback version is used, the
gain of the triode can be tailored to suit the rest of the circuit.

Good idea !! I'll look into that mixer thing again !!

But remember such a stage inverts the phase, and perhaps you should try to
have
phase coherence between bass and sub signals.

I thought of that .......

And since the phase tunrs in such a mixer stage I think it's better to
take
the L and R signal from the cathodes of the 5687 .
Loosing the gain is more like a pro than a con !!

See ? ;-))
And a phase flip switch on a sub is always a good idea I think .
( or just swap the wires on the speaker ....)
If you turn the sub 180 dgr. and move it to the other side of the room .....
Ok , better not , but if you can't do it the way it should , it should be
the way it can ..... ;-))

You may need some gain since you will need to have a separate bass
level control to get the level of sub-bass about right with the rest of
the
sound.

No problem . I planned that on the sub somehow .

The signal could be summed via 100k resistors from each anode
of the 5687 and the junction of these R can have a cap to ground with
a pole at about 200Hz.
The R junction is also taken to a single CF to make a buffer before the
second order filter with pole at say 150, and you will have 18 dB/octave
slope above 300 Hz.

I don't see the need of that 18dB/octave slope , 12dB/octave should do
the trick since the satelites will (roughly) have a 12dB/octave slope too .
A closed box (Qtc = 0.707) has a 6dB/octave slope and the plan was
to tune the HP filter (upper part schematic) at the same Fc , giving a
12dB/otave Linkwitz-Riley slope (Q = 0.5) for the satelites .

The only thing missing is a gain control, but the level of bass recoverd
as suggested will be higher that fed to your PL86.

Yes , roughly double . No big problem . I planned a TDA2030 chip(s) for
the sub-amp , bridged and/or boosted with a pair of transistors ......
A soundcard can drive that chip to full output , so I don't need extra gain
I
think and so it might be better to take the signal for the mixer from the
cathode of the 5687
And if so , what harm can an extra op-amp do if I use the TDA2030 anyway .
Maybe a PP EL34 or 6L6 in the future , but I think a TDA2030 will do just
fine for the rumbling in the sub .....

But for a passive RC type with fairly rapid roll off with ultimate
attenuation of 12 dB/octave, you can have a low impedance source
such as a cathode follower driving 10k with 0.068 uF to 0V,
giving a pole at 234Hz.
Then a following RC section with 47k and 0.015 uF for a pole at 212 Hz.
The second RC filter has 5 times the impedance value of the first,
so they act to give a reasonably sharper roll off than if you had two RC
filters

with the same values.
The combined -3dB pole will be at around 150 Hz, and
by 500 Hz the slope has become steep.
Its naturally damped, and simple.

Hmm , very odd dB slopes , way to close to Fc , giving way to odd phase
shifts I think .
The 12dB/octave Linkwitz-Riley LP filter shown in my schematic can simply
be tuned (vary Fc) by turning the dual 100k pot .


Yes, but research into sub-woofer powered sports cars with solar roof
panels
is underway.

They are training the bass player and drummer to provide the right
type of sound to use, and design of suitable ear muffs for the driver
is just about done.
One British horn speaker maker was contracted to provide a horn powered
protoype,
but they had trouble getting a permit from the transport authorities,
since it couldn't fit under the bridges.
Then someone said it'd be more effective to use sub powered barges, and
limit
them to use in countryside canals, but the Society Against Hearing Damage
in
Tortoises
gained an injunction in the High Court against underwater subs
and underwater horns.

Progress continues in the face of obstacles.

LOL !!!!

Now I understand why Michael Schumachel already has his own brand of
sub-woofers !!!!

Ronald .

Patrick Turner wrote:

Ronald wrote:


Hi Patrick ,


Try Duncans amp pages, go tube data, there are a few links.....

I found and saved a pdf 2 nights ago when looking for
E184F/6EJ7 pentode data.
I also found EF80/6BX6.

I did , but didn't find exact data on triode connection .

Patrick/Ronald, there's a set of triode curves on page 6 of the data
sheet for the EF80 http://tdsl.duncanamps.com/link.php?target=00254B8C

--
Duncan Munro
http://www.duncanamps.com/

Hi Duncan ,

Patrick/Ronald, there's a set of triode curves on page 6 of the data
sheet for the EF80 http://tdsl.duncanamps.com/link.php?target=00254B8C

--
Duncan Munro
http://www.duncanamps.com/

That'll do for a CF , but it's not the type of curve I'm used to .
It's different from the curves you'll see at "real" triodes .
(Ua on X-scale , Ia on Y-scale) .
Maybe I have to learn to read "odd" curves better ...... (?)
I'm a bit puzzled about using this curve for a grounded cathode stage ....

Thanks ,

Ronald .

P.s.
Many times I take a "shortcut" to Frank's but still I use your site and
proggies a lot !!
So keep up the good work !! PSU designer is great !!

Ronald wrote:

Hi Pat ,

OK, nor could I and some of the links from Duncan's pages don't work.

Most of the time I go directly to Frank's .....
http://www.mif.pg.gda.pl/homepages/frank/
There are many tube not linked to Duncan's !!

The internal amplificayion factor from g1 g2 is given as 50.
I measured 6BX6 with triode µ =50, Ra around 12k.

That would give a S = 4.16 mA/V , indeed pretty close to a 12AT7 and
at least good enough for a CF .

The bjts are TO26 types and about $2 each from most outlets.
Its a very cheap and effective solution.

Ok , $2 on a entire project is peanuts .....

I saw something like that at the Tube Cad Journal , mixing L+R and use
the same R from the anode back to the grid .
http://www.tubecad.com/april_may2001/page22.html

The tube cad method is ok because the LR signals are summed outside the
triode, and applied to the grid as one signal.
Then if the shunt feedback version is used, the
gain of the triode can be tailored to suit the rest of the circuit.

Good idea !! I'll look into that mixer thing again !!

But remember such a stage inverts the phase, and perhaps you should try to
have
phase coherence between bass and sub signals.

I thought of that .......

And since the phase tunrs in such a mixer stage I think it's better to
take
the L and R signal from the cathodes of the 5687 .
Loosing the gain is more like a pro than a con !!

See ? ;-))
And a phase flip switch on a sub is always a good idea I think .
( or just swap the wires on the speaker ....)
If you turn the sub 180 dgr. and move it to the other side of the room .....
Ok , better not , but if you can't do it the way it should , it should be
the way it can ..... ;-))

You may need some gain since you will need to have a separate bass
level control to get the level of sub-bass about right with the rest of
the
sound.

No problem . I planned that on the sub somehow .

The signal could be summed via 100k resistors from each anode
of the 5687 and the junction of these R can have a cap to ground with
a pole at about 200Hz.
The R junction is also taken to a single CF to make a buffer before the
second order filter with pole at say 150, and you will have 18 dB/octave
slope above 300 Hz.

I don't see the need of that 18dB/octave slope , 12dB/octave should do
the trick since the satelites will (roughly) have a 12dB/octave slope too .
A closed box (Qtc = 0.707) has a 6dB/octave slope and the plan was
to tune the HP filter (upper part schematic) at the same Fc , giving a
12dB/otave Linkwitz-Riley slope (Q = 0.5) for the satelites .

Frequencies above 100 Hz become more noticeable to the ear
as F rises. The large cone area of a sub makes the F between say 150 and say 1k
far too noticeable unless you filter them well, hence my use of
a filter with a 24 dB/octave slope by 500 Hz.

The slopes of filters on the satelites don't matter as much and
just to remove their low bass a large series cap may do
for a first order filter.
Satelites make little deep bass, and the bass they do make become less noticable
as F falls.
Some overlap is ok.





The only thing missing is a gain control, but the level of bass recoverd
as suggested will be higher that fed to your PL86.

Yes , roughly double . No big problem . I planned a TDA2030 chip(s) for
the sub-amp , bridged and/or boosted with a pair of transistors ......
A soundcard can drive that chip to full output , so I don't need extra gain
I
think and so it might be better to take the signal for the mixer from the
cathode of the 5687
And if so , what harm can an extra op-amp do if I use the TDA2030 anyway .
Maybe a PP EL34 or 6L6 in the future , but I think a TDA2030 will do just
fine for the rumbling in the sub .....

But for a passive RC type with fairly rapid roll off with ultimate
attenuation of 12 dB/octave, you can have a low impedance source
such as a cathode follower driving 10k with 0.068 uF to 0V,
giving a pole at 234Hz.
Then a following RC section with 47k and 0.015 uF for a pole at 212 Hz.
The second RC filter has 5 times the impedance value of the first,
so they act to give a reasonably sharper roll off than if you had two RC
filters

with the same values.
The combined -3dB pole will be at around 150 Hz, and
by 500 Hz the slope has become steep.
Its naturally damped, and simple.

Hmm , very odd dB slopes , way to close to Fc , giving way to odd phase
shifts I think .

Nope, the passive RC filters have gentle phase shift at the Fc.
A second order filter se to make a flat response followed by as steep
as possible slope of 12 dB makes 90 degress of phase shift at -3dB.
And a 24dB /octave similarly "critically damped" filter has 108d shift at
its -3dBb Fc.




The 12dB/octave Linkwitz-Riley LP filter shown in my schematic can simply
be tuned (vary Fc) by turning the dual 100k pot .

What happens to the shape of the response curve?

With passive RC, I use a 3 position switch to change from
say 37, 55 and 90 Hz cut off points by switching the value of C.

No multiway pots are needed.


Yes, but research into sub-woofer powered sports cars with solar roof
panels
is underway.

They are training the bass player and drummer to provide the right
type of sound to use, and design of suitable ear muffs for the driver
is just about done.
One British horn speaker maker was contracted to provide a horn powered
protoype,
but they had trouble getting a permit from the transport authorities,
since it couldn't fit under the bridges.
Then someone said it'd be more effective to use sub powered barges, and
limit
them to use in countryside canals, but the Society Against Hearing Damage
in
Tortoises
gained an injunction in the High Court against underwater subs
and underwater horns.

Progress continues in the face of obstacles.

LOL !!!!

Now I understand why Michael Schumachel already has his own brand of
sub-woofers !!!!

Ah, there's money to be made if you get famous and faster than everyone else.
I wish him well and hope he don't do an Ayton Senna Barrier Interface
Experiment.

Patrick Turner.



Ronald .

Duncan Munro wrote:

Patrick Turner wrote:

Ronald wrote:


Hi Patrick ,


Try Duncans amp pages, go tube data, there are a few links.....

I found and saved a pdf 2 nights ago when looking for
E184F/6EJ7 pentode data.
I also found EF80/6BX6.

I did , but didn't find exact data on triode connection .

Patrick/Ronald, there's a set of triode curves on page 6 of the data
sheet for the EF80 http://tdsl.duncanamps.com/link.php?target=00254B8C

Unfortunately, the "triode curves" you refer to are not
triode curves; they just look like triode curves, since they are
gm curves for the pentode operation.

There are no Ia x Ea triode curves.

But the 6BX6 / EF80 makes a fine triode.

Patrick Turner.



--
Duncan Munro
http://www.duncanamps.com/

Hi Pat ,

I think I'll stick to my first plan for the amp / filtering inside the PC .
I made the chassis / frame for the amp yesterday and it's gonna be
tight already I think .
The TDA2030 basicly is an opamp that can deliver 14W into 4 ohms .
http://www.ben.cz/download/datasheet/tda2030a.pdf
(TDA2030 = 14W , TDA2030A = 18W . I have the 14W)
So if needed I can add some extra filtering there or even add an extra
opamp . Who cares , it's a sand amp anyway ;-))

Thanks for all the help !!

Ronald .


"Patrick Turner" schreef in bericht
...


Ronald wrote:

Hi Pat ,

OK, nor could I and some of the links from Duncan's pages don't work.

Most of the time I go directly to Frank's .....
http://www.mif.pg.gda.pl/homepages/frank/
There are many tube not linked to Duncan's !!

The internal amplificayion factor from g1 g2 is given as 50.
I measured 6BX6 with triode µ =50, Ra around 12k.

That would give a S = 4.16 mA/V , indeed pretty close to a 12AT7 and
at least good enough for a CF .

The bjts are TO26 types and about $2 each from most outlets.
Its a very cheap and effective solution.

Ok , $2 on a entire project is peanuts .....

I saw something like that at the Tube Cad Journal , mixing L+R and
use
the same R from the anode back to the grid .
http://www.tubecad.com/april_may2001/page22.html

The tube cad method is ok because the LR signals are summed outside
the
triode, and applied to the grid as one signal.
Then if the shunt feedback version is used, the
gain of the triode can be tailored to suit the rest of the circuit.

Good idea !! I'll look into that mixer thing again !!

But remember such a stage inverts the phase, and perhaps you should
try to
have
phase coherence between bass and sub signals.

I thought of that .......

And since the phase tunrs in such a mixer stage I think it's better
to
take
the L and R signal from the cathodes of the 5687 .
Loosing the gain is more like a pro than a con !!

See ? ;-))
And a phase flip switch on a sub is always a good idea I think .
( or just swap the wires on the speaker ....)
If you turn the sub 180 dgr. and move it to the other side of the room
......
Ok , better not , but if you can't do it the way it should , it should
be
the way it can ..... ;-))

You may need some gain since you will need to have a separate bass
level control to get the level of sub-bass about right with the rest
of
the
sound.

No problem . I planned that on the sub somehow .

The signal could be summed via 100k resistors from each anode
of the 5687 and the junction of these R can have a cap to ground with
a pole at about 200Hz.
The R junction is also taken to a single CF to make a buffer before
the
second order filter with pole at say 150, and you will have 18
dB/octave
slope above 300 Hz.

I don't see the need of that 18dB/octave slope , 12dB/octave should do
the trick since the satelites will (roughly) have a 12dB/octave slope
too .
A closed box (Qtc = 0.707) has a 6dB/octave slope and the plan was
to tune the HP filter (upper part schematic) at the same Fc , giving a
12dB/otave Linkwitz-Riley slope (Q = 0.5) for the satelites .

Frequencies above 100 Hz become more noticeable to the ear
as F rises. The large cone area of a sub makes the F between say 150 and
say 1k
far too noticeable unless you filter them well, hence my use of
a filter with a 24 dB/octave slope by 500 Hz.

The slopes of filters on the satelites don't matter as much and
just to remove their low bass a large series cap may do
for a first order filter.
Satelites make little deep bass, and the bass they do make become less
noticable
as F falls.
Some overlap is ok.





The only thing missing is a gain control, but the level of bass
recoverd
as suggested will be higher that fed to your PL86.

Yes , roughly double . No big problem . I planned a TDA2030 chip(s) for
the sub-amp , bridged and/or boosted with a pair of transistors ......
A soundcard can drive that chip to full output , so I don't need extra
gain
I
think and so it might be better to take the signal for the mixer from
the
cathode of the 5687
And if so , what harm can an extra op-amp do if I use the TDA2030 anyway
..
Maybe a PP EL34 or 6L6 in the future , but I think a TDA2030 will do
just
fine for the rumbling in the sub .....

But for a passive RC type with fairly rapid roll off with ultimate
attenuation of 12 dB/octave, you can have a low impedance source
such as a cathode follower driving 10k with 0.068 uF to 0V,
giving a pole at 234Hz.
Then a following RC section with 47k and 0.015 uF for a pole at 212
Hz.
The second RC filter has 5 times the impedance value of the first,
so they act to give a reasonably sharper roll off than if you had two
RC
filters

with the same values.
The combined -3dB pole will be at around 150 Hz, and
by 500 Hz the slope has become steep.
Its naturally damped, and simple.

Hmm , very odd dB slopes , way to close to Fc , giving way to odd phase
shifts I think .

Nope, the passive RC filters have gentle phase shift at the Fc.
A second order filter se to make a flat response followed by as steep
as possible slope of 12 dB makes 90 degress of phase shift at -3dB.
And a 24dB /octave similarly "critically damped" filter has 108d shift at
its -3dBb Fc.




The 12dB/octave Linkwitz-Riley LP filter shown in my schematic can
simply
be tuned (vary Fc) by turning the dual 100k pot .

What happens to the shape of the response curve?

With passive RC, I use a 3 position switch to change from
say 37, 55 and 90 Hz cut off points by switching the value of C.

No multiway pots are needed.


Yes, but research into sub-woofer powered sports cars with solar roof
panels
is underway.

They are training the bass player and drummer to provide the right
type of sound to use, and design of suitable ear muffs for the driver
is just about done.
One British horn speaker maker was contracted to provide a horn
powered
protoype,
but they had trouble getting a permit from the transport authorities,
since it couldn't fit under the bridges.
Then someone said it'd be more effective to use sub powered barges,
and
limit
them to use in countryside canals, but the Society Against Hearing
Damage
in
Tortoises
gained an injunction in the High Court against underwater subs
and underwater horns.

Progress continues in the face of obstacles.

LOL !!!!

Now I understand why Michael Schumachel already has his own brand of
sub-woofers !!!!

Ah, there's money to be made if you get famous and faster than everyone
else.
I wish him well and hope he don't do an Ayton Senna Barrier Interface
Experiment.

Patrick Turner.



Ronald .

Ronald wrote:

Hi Pat ,

I think I'll stick to my first plan for the amp / filtering inside the PC .
I made the chassis / frame for the amp yesterday and it's gonna be
tight already I think .
The TDA2030 basicly is an opamp that can deliver 14W into 4 ohms .
http://www.ben.cz/download/datasheet/tda2030a.pdf
(TDA2030 = 14W , TDA2030A = 18W . I have the 14W)
So if needed I can add some extra filtering there or even add an extra
opamp . Who cares , it's a sand amp anyway ;-))

Make sure you have sufficient heatsinking for the amp chips.

Patrick Turner.



Thanks for all the help !!

Ronald .

"Patrick Turner" schreef in bericht
...


Ronald wrote:

Hi Pat ,

OK, nor could I and some of the links from Duncan's pages don't work.

Most of the time I go directly to Frank's .....
http://www.mif.pg.gda.pl/homepages/frank/
There are many tube not linked to Duncan's !!

The internal amplificayion factor from g1 g2 is given as 50.
I measured 6BX6 with triode µ =50, Ra around 12k.

That would give a S = 4.16 mA/V , indeed pretty close to a 12AT7 and
at least good enough for a CF .

The bjts are TO26 types and about $2 each from most outlets.
Its a very cheap and effective solution.

Ok , $2 on a entire project is peanuts .....

I saw something like that at the Tube Cad Journal , mixing L+R and
use
the same R from the anode back to the grid .
http://www.tubecad.com/april_may2001/page22.html

The tube cad method is ok because the LR signals are summed outside
the
triode, and applied to the grid as one signal.
Then if the shunt feedback version is used, the
gain of the triode can be tailored to suit the rest of the circuit.

Good idea !! I'll look into that mixer thing again !!

But remember such a stage inverts the phase, and perhaps you should
try to
have
phase coherence between bass and sub signals.

I thought of that .......

And since the phase tunrs in such a mixer stage I think it's better
to
take
the L and R signal from the cathodes of the 5687 .
Loosing the gain is more like a pro than a con !!

See ? ;-))
And a phase flip switch on a sub is always a good idea I think .
( or just swap the wires on the speaker ....)
If you turn the sub 180 dgr. and move it to the other side of the room
.....
Ok , better not , but if you can't do it the way it should , it should
be
the way it can ..... ;-))

You may need some gain since you will need to have a separate bass
level control to get the level of sub-bass about right with the rest
of
the
sound.

No problem . I planned that on the sub somehow .

The signal could be summed via 100k resistors from each anode
of the 5687 and the junction of these R can have a cap to ground with
a pole at about 200Hz.
The R junction is also taken to a single CF to make a buffer before
the
second order filter with pole at say 150, and you will have 18
dB/octave
slope above 300 Hz.

I don't see the need of that 18dB/octave slope , 12dB/octave should do
the trick since the satelites will (roughly) have a 12dB/octave slope
too .
A closed box (Qtc = 0.707) has a 6dB/octave slope and the plan was
to tune the HP filter (upper part schematic) at the same Fc , giving a
12dB/otave Linkwitz-Riley slope (Q = 0.5) for the satelites .

Frequencies above 100 Hz become more noticeable to the ear
as F rises. The large cone area of a sub makes the F between say 150 and
say 1k
far too noticeable unless you filter them well, hence my use of
a filter with a 24 dB/octave slope by 500 Hz.

The slopes of filters on the satelites don't matter as much and
just to remove their low bass a large series cap may do
for a first order filter.
Satelites make little deep bass, and the bass they do make become less
noticable
as F falls.
Some overlap is ok.





The only thing missing is a gain control, but the level of bass
recoverd
as suggested will be higher that fed to your PL86.

Yes , roughly double . No big problem . I planned a TDA2030 chip(s) for
the sub-amp , bridged and/or boosted with a pair of transistors ......
A soundcard can drive that chip to full output , so I don't need extra
gain
I
think and so it might be better to take the signal for the mixer from
the
cathode of the 5687
And if so , what harm can an extra op-amp do if I use the TDA2030 anyway
.
Maybe a PP EL34 or 6L6 in the future , but I think a TDA2030 will do
just
fine for the rumbling in the sub .....

But for a passive RC type with fairly rapid roll off with ultimate
attenuation of 12 dB/octave, you can have a low impedance source
such as a cathode follower driving 10k with 0.068 uF to 0V,
giving a pole at 234Hz.
Then a following RC section with 47k and 0.015 uF for a pole at 212
Hz.
The second RC filter has 5 times the impedance value of the first,
so they act to give a reasonably sharper roll off than if you had two
RC
filters

with the same values.
The combined -3dB pole will be at around 150 Hz, and
by 500 Hz the slope has become steep.
Its naturally damped, and simple.

Hmm , very odd dB slopes , way to close to Fc , giving way to odd phase
shifts I think .

Nope, the passive RC filters have gentle phase shift at the Fc.
A second order filter se to make a flat response followed by as steep
as possible slope of 12 dB makes 90 degress of phase shift at -3dB.
And a 24dB /octave similarly "critically damped" filter has 108d shift at
its -3dBb Fc.




The 12dB/octave Linkwitz-Riley LP filter shown in my schematic can
simply
be tuned (vary Fc) by turning the dual 100k pot .

What happens to the shape of the response curve?

With passive RC, I use a 3 position switch to change from
say 37, 55 and 90 Hz cut off points by switching the value of C.

No multiway pots are needed.


Yes, but research into sub-woofer powered sports cars with solar roof
panels
is underway.

They are training the bass player and drummer to provide the right
type of sound to use, and design of suitable ear muffs for the driver
is just about done.
One British horn speaker maker was contracted to provide a horn
powered
protoype,
but they had trouble getting a permit from the transport authorities,
since it couldn't fit under the bridges.
Then someone said it'd be more effective to use sub powered barges,
and
limit
them to use in countryside canals, but the Society Against Hearing
Damage
in
Tortoises
gained an injunction in the High Court against underwater subs
and underwater horns.

Progress continues in the face of obstacles.

LOL !!!!

Now I understand why Michael Schumachel already has his own brand of
sub-woofers !!!!

Ah, there's money to be made if you get famous and faster than everyone
else.
I wish him well and hope he don't do an Ayton Senna Barrier Interface
Experiment.

Patrick Turner.



Ronald .

Hi Pat ,

I have a nice heatsink from a 2x 40W amp so if that doesn't
do the trick , what does ... ;-))
The same sink with a fan I think . The sub is for PC use only
so I think I don't care about an extra "vacuüm cleaner" ;-))

Thanks for the warning anyway ,

Ronald .

"Patrick Turner" schreef in bericht
...


Ronald wrote:

Hi Pat ,

I think I'll stick to my first plan for the amp / filtering inside the
PC .
I made the chassis / frame for the amp yesterday and it's gonna be
tight already I think .
The TDA2030 basicly is an opamp that can deliver 14W into 4 ohms .
http://www.ben.cz/download/datasheet/tda2030a.pdf
(TDA2030 = 14W , TDA2030A = 18W . I have the 14W)
So if needed I can add some extra filtering there or even add an extra
opamp . Who cares , it's a sand amp anyway ;-))

Make sure you have sufficient heatsinking for the amp chips.

Patrick Turner.



Thanks for all the help !!

Ronald .

"Patrick Turner" schreef in bericht
...


Ronald wrote:

Hi Pat ,

OK, nor could I and some of the links from Duncan's pages don't
work.

Most of the time I go directly to Frank's .....
http://www.mif.pg.gda.pl/homepages/frank/
There are many tube not linked to Duncan's !!

The internal amplificayion factor from g1 g2 is given as 50.
I measured 6BX6 with triode µ =50, Ra around 12k.

That would give a S = 4.16 mA/V , indeed pretty close to a 12AT7 and
at least good enough for a CF .

The bjts are TO26 types and about $2 each from most outlets.
Its a very cheap and effective solution.

Ok , $2 on a entire project is peanuts .....

I saw something like that at the Tube Cad Journal , mixing L+R
and
use
the same R from the anode back to the grid .
http://www.tubecad.com/april_may2001/page22.html

The tube cad method is ok because the LR signals are summed
outside
the
triode, and applied to the grid as one signal.
Then if the shunt feedback version is used, the
gain of the triode can be tailored to suit the rest of the
circuit.

Good idea !! I'll look into that mixer thing again !!

But remember such a stage inverts the phase, and perhaps you
should
try to
have
phase coherence between bass and sub signals.

I thought of that .......

And since the phase tunrs in such a mixer stage I think it's
better
to
take
the L and R signal from the cathodes of the 5687 .
Loosing the gain is more like a pro than a con !!

See ? ;-))
And a phase flip switch on a sub is always a good idea I think .
( or just swap the wires on the speaker ....)
If you turn the sub 180 dgr. and move it to the other side of the
room
.....
Ok , better not , but if you can't do it the way it should , it
should
be
the way it can ..... ;-))

You may need some gain since you will need to have a separate bass
level control to get the level of sub-bass about right with the
rest
of
the
sound.

No problem . I planned that on the sub somehow .

The signal could be summed via 100k resistors from each anode
of the 5687 and the junction of these R can have a cap to ground
with
a pole at about 200Hz.
The R junction is also taken to a single CF to make a buffer
before
the
second order filter with pole at say 150, and you will have 18
dB/octave
slope above 300 Hz.

I don't see the need of that 18dB/octave slope , 12dB/octave should
do
the trick since the satelites will (roughly) have a 12dB/octave
slope
too .
A closed box (Qtc = 0.707) has a 6dB/octave slope and the plan was
to tune the HP filter (upper part schematic) at the same Fc , giving
a
12dB/otave Linkwitz-Riley slope (Q = 0.5) for the satelites .

Frequencies above 100 Hz become more noticeable to the ear
as F rises. The large cone area of a sub makes the F between say 150
and
say 1k
far too noticeable unless you filter them well, hence my use of
a filter with a 24 dB/octave slope by 500 Hz.

The slopes of filters on the satelites don't matter as much and
just to remove their low bass a large series cap may do
for a first order filter.
Satelites make little deep bass, and the bass they do make become less
noticable
as F falls.
Some overlap is ok.





The only thing missing is a gain control, but the level of bass
recoverd
as suggested will be higher that fed to your PL86.

Yes , roughly double . No big problem . I planned a TDA2030 chip(s)
for
the sub-amp , bridged and/or boosted with a pair of transistors
.......
A soundcard can drive that chip to full output , so I don't need
extra
gain
I
think and so it might be better to take the signal for the mixer
from
the
cathode of the 5687
And if so , what harm can an extra op-amp do if I use the TDA2030
anyway
.
Maybe a PP EL34 or 6L6 in the future , but I think a TDA2030 will do
just
fine for the rumbling in the sub .....

But for a passive RC type with fairly rapid roll off with ultimate
attenuation of 12 dB/octave, you can have a low impedance source
such as a cathode follower driving 10k with 0.068 uF to 0V,
giving a pole at 234Hz.
Then a following RC section with 47k and 0.015 uF for a pole at
212
Hz.
The second RC filter has 5 times the impedance value of the first,
so they act to give a reasonably sharper roll off than if you had
two
RC
filters

with the same values.
The combined -3dB pole will be at around 150 Hz, and
by 500 Hz the slope has become steep.
Its naturally damped, and simple.

Hmm , very odd dB slopes , way to close to Fc , giving way to odd
phase
shifts I think .

Nope, the passive RC filters have gentle phase shift at the Fc.
A second order filter se to make a flat response followed by as steep
as possible slope of 12 dB makes 90 degress of phase shift at -3dB.
And a 24dB /octave similarly "critically damped" filter has 108d
shift at
its -3dBb Fc.




The 12dB/octave Linkwitz-Riley LP filter shown in my schematic can
simply
be tuned (vary Fc) by turning the dual 100k pot .

What happens to the shape of the response curve?

With passive RC, I use a 3 position switch to change from
say 37, 55 and 90 Hz cut off points by switching the value of C.

No multiway pots are needed.


Yes, but research into sub-woofer powered sports cars with solar
roof
panels
is underway.

They are training the bass player and drummer to provide the right
type of sound to use, and design of suitable ear muffs for the
driver
is just about done.
One British horn speaker maker was contracted to provide a horn
powered
protoype,
but they had trouble getting a permit from the transport
authorities,
since it couldn't fit under the bridges.
Then someone said it'd be more effective to use sub powered
barges,
and
limit
them to use in countryside canals, but the Society Against Hearing
Damage
in
Tortoises
gained an injunction in the High Court against underwater subs
and underwater horns.

Progress continues in the face of obstacles.

LOL !!!!

Now I understand why Michael Schumachel already has his own brand of
sub-woofers !!!!

Ah, there's money to be made if you get famous and faster than
everyone
else.
I wish him well and hope he don't do an Ayton Senna Barrier Interface
Experiment.

Patrick Turner.



Ronald .

Ronald wrote:

Hi Pat ,

I have a nice heatsink from a 2x 40W amp so if that doesn't
do the trick , what does ... ;-))
The same sink with a fan I think . The sub is for PC use only
so I think I don't care about an extra "vacuüm cleaner" ;-))

Thanks for the warning anyway ,

Ronald .

Things can get warm inside a PC case in summer.
I hate the darn fan I have in my 8 yr old PC with a 166 Mhz processor,
and I have yet to get around to fitting a quiet fan.
An amp heatsink just needs a good slow draft past it, so a slow big fan
is OK, maybe you have that already.

Patrick Turner.




"Patrick Turner" schreef in bericht
...


Ronald wrote:

Hi Pat ,

I think I'll stick to my first plan for the amp / filtering inside the
PC .
I made the chassis / frame for the amp yesterday and it's gonna be
tight already I think .
The TDA2030 basicly is an opamp that can deliver 14W into 4 ohms .
http://www.ben.cz/download/datasheet/tda2030a.pdf
(TDA2030 = 14W , TDA2030A = 18W . I have the 14W)
So if needed I can add some extra filtering there or even add an extra
opamp . Who cares , it's a sand amp anyway ;-))

Make sure you have sufficient heatsinking for the amp chips.

Patrick Turner.



Thanks for all the help !!

Ronald .

"Patrick Turner" schreef in bericht
...


Ronald wrote:

Hi Pat ,

OK, nor could I and some of the links from Duncan's pages don't
work.

Most of the time I go directly to Frank's .....
http://www.mif.pg.gda.pl/homepages/frank/
There are many tube not linked to Duncan's !!

The internal amplificayion factor from g1 g2 is given as 50.
I measured 6BX6 with triode µ =50, Ra around 12k.

That would give a S = 4.16 mA/V , indeed pretty close to a 12AT7 and
at least good enough for a CF .

The bjts are TO26 types and about $2 each from most outlets.
Its a very cheap and effective solution.

Ok , $2 on a entire project is peanuts .....

I saw something like that at the Tube Cad Journal , mixing L+R
and
use
the same R from the anode back to the grid .
http://www.tubecad.com/april_may2001/page22.html

The tube cad method is ok because the LR signals are summed
outside
the
triode, and applied to the grid as one signal.
Then if the shunt feedback version is used, the
gain of the triode can be tailored to suit the rest of the
circuit.

Good idea !! I'll look into that mixer thing again !!

But remember such a stage inverts the phase, and perhaps you
should
try to
have
phase coherence between bass and sub signals.

I thought of that .......

And since the phase tunrs in such a mixer stage I think it's
better
to
take
the L and R signal from the cathodes of the 5687 .
Loosing the gain is more like a pro than a con !!

See ? ;-))
And a phase flip switch on a sub is always a good idea I think .
( or just swap the wires on the speaker ....)
If you turn the sub 180 dgr. and move it to the other side of the
room
.....
Ok , better not , but if you can't do it the way it should , it
should
be
the way it can ..... ;-))

You may need some gain since you will need to have a separate bass
level control to get the level of sub-bass about right with the
rest
of
the
sound.

No problem . I planned that on the sub somehow .

The signal could be summed via 100k resistors from each anode
of the 5687 and the junction of these R can have a cap to ground
with
a pole at about 200Hz.
The R junction is also taken to a single CF to make a buffer
before
the
second order filter with pole at say 150, and you will have 18
dB/octave
slope above 300 Hz.

I don't see the need of that 18dB/octave slope , 12dB/octave should
do
the trick since the satelites will (roughly) have a 12dB/octave
slope
too .
A closed box (Qtc = 0.707) has a 6dB/octave slope and the plan was
to tune the HP filter (upper part schematic) at the same Fc , giving
a
12dB/otave Linkwitz-Riley slope (Q = 0.5) for the satelites .

Frequencies above 100 Hz become more noticeable to the ear
as F rises. The large cone area of a sub makes the F between say 150
and
say 1k
far too noticeable unless you filter them well, hence my use of
a filter with a 24 dB/octave slope by 500 Hz.

The slopes of filters on the satelites don't matter as much and
just to remove their low bass a large series cap may do
for a first order filter.
Satelites make little deep bass, and the bass they do make become less
noticable
as F falls.
Some overlap is ok.





The only thing missing is a gain control, but the level of bass
recoverd
as suggested will be higher that fed to your PL86.

Yes , roughly double . No big problem . I planned a TDA2030 chip(s)
for
the sub-amp , bridged and/or boosted with a pair of transistors
......
A soundcard can drive that chip to full output , so I don't need
extra
gain
I
think and so it might be better to take the signal for the mixer
from
the
cathode of the 5687
And if so , what harm can an extra op-amp do if I use the TDA2030
anyway
.
Maybe a PP EL34 or 6L6 in the future , but I think a TDA2030 will do
just
fine for the rumbling in the sub .....

But for a passive RC type with fairly rapid roll off with ultimate
attenuation of 12 dB/octave, you can have a low impedance source
such as a cathode follower driving 10k with 0.068 uF to 0V,
giving a pole at 234Hz.
Then a following RC section with 47k and 0.015 uF for a pole at
212
Hz.
The second RC filter has 5 times the impedance value of the first,
so they act to give a reasonably sharper roll off than if you had
two
RC
filters

with the same values.
The combined -3dB pole will be at around 150 Hz, and
by 500 Hz the slope has become steep.
Its naturally damped, and simple.

Hmm , very odd dB slopes , way to close to Fc , giving way to odd
phase
shifts I think .

Nope, the passive RC filters have gentle phase shift at the Fc.
A second order filter se to make a flat response followed by as steep
as possible slope of 12 dB makes 90 degress of phase shift at -3dB.
And a 24dB /octave similarly "critically damped" filter has 108d
shift at
its -3dBb Fc.




The 12dB/octave Linkwitz-Riley LP filter shown in my schematic can
simply
be tuned (vary Fc) by turning the dual 100k pot .

What happens to the shape of the response curve?

With passive RC, I use a 3 position switch to change from
say 37, 55 and 90 Hz cut off points by switching the value of C.

No multiway pots are needed.


Yes, but research into sub-woofer powered sports cars with solar
roof
panels
is underway.

They are training the bass player and drummer to provide the right
type of sound to use, and design of suitable ear muffs for the
driver
is just about done.
One British horn speaker maker was contracted to provide a horn
powered
protoype,
but they had trouble getting a permit from the transport
authorities,
since it couldn't fit under the bridges.
Then someone said it'd be more effective to use sub powered
barges,
and
limit
them to use in countryside canals, but the Society Against Hearing
Damage
in
Tortoises
gained an injunction in the High Court against underwater subs
and underwater horns.

Progress continues in the face of obstacles.

LOL !!!!

Now I understand why Michael Schumachel already has his own brand of
sub-woofers !!!!

Ah, there's money to be made if you get famous and faster than
everyone
else.
I wish him well and hope he don't do an Ayton Senna Barrier Interface
Experiment.

Patrick Turner.



Ronald .

Hi Patrick ,

I was planning the sub-amp on the back of the sub .
No need to show a TDA2030 behind that "window" if
there's a AMD 2500 XP already ;-))

Ronald .


"Patrick Turner" schreef in bericht
...


Ronald wrote:

Hi Pat ,

I have a nice heatsink from a 2x 40W amp so if that doesn't
do the trick , what does ... ;-))
The same sink with a fan I think . The sub is for PC use only
so I think I don't care about an extra "vacuüm cleaner" ;-))

Thanks for the warning anyway ,

Ronald .

Things can get warm inside a PC case in summer.
I hate the darn fan I have in my 8 yr old PC with a 166 Mhz processor,
and I have yet to get around to fitting a quiet fan.
An amp heatsink just needs a good slow draft past it, so a slow big fan
is OK, maybe you have that already.

Patrick Turner.




"Patrick Turner" schreef in bericht
...


Ronald wrote:

Hi Pat ,

I think I'll stick to my first plan for the amp / filtering inside
the
PC .
I made the chassis / frame for the amp yesterday and it's gonna be
tight already I think .
The TDA2030 basicly is an opamp that can deliver 14W into 4 ohms .
http://www.ben.cz/download/datasheet/tda2030a.pdf
(TDA2030 = 14W , TDA2030A = 18W . I have the 14W)
So if needed I can add some extra filtering there or even add an
extra
opamp . Who cares , it's a sand amp anyway ;-))

Make sure you have sufficient heatsinking for the amp chips.

Patrick Turner.



Thanks for all the help !!

Ronald .

"Patrick Turner" schreef in bericht
...


Ronald wrote:

Hi Pat ,

OK, nor could I and some of the links from Duncan's pages
don't
work.

Most of the time I go directly to Frank's .....
http://www.mif.pg.gda.pl/homepages/frank/
There are many tube not linked to Duncan's !!

The internal amplificayion factor from g1 g2 is given as 50.
I measured 6BX6 with triode µ =50, Ra around 12k.

That would give a S = 4.16 mA/V , indeed pretty close to a 12AT7
and
at least good enough for a CF .

The bjts are TO26 types and about $2 each from most outlets.
Its a very cheap and effective solution.

Ok , $2 on a entire project is peanuts .....

I saw something like that at the Tube Cad Journal , mixing
L+R
and
use
the same R from the anode back to the grid .
http://www.tubecad.com/april_may2001/page22.html

The tube cad method is ok because the LR signals are summed
outside
the
triode, and applied to the grid as one signal.
Then if the shunt feedback version is used, the
gain of the triode can be tailored to suit the rest of the
circuit.

Good idea !! I'll look into that mixer thing again !!

But remember such a stage inverts the phase, and perhaps you
should
try to
have
phase coherence between bass and sub signals.

I thought of that .......

And since the phase tunrs in such a mixer stage I think it's
better
to
take
the L and R signal from the cathodes of the 5687 .
Loosing the gain is more like a pro than a con !!

See ? ;-))
And a phase flip switch on a sub is always a good idea I think .
( or just swap the wires on the speaker ....)
If you turn the sub 180 dgr. and move it to the other side of
the
room
.....
Ok , better not , but if you can't do it the way it should , it
should
be
the way it can ..... ;-))

You may need some gain since you will need to have a separate
bass
level control to get the level of sub-bass about right with
the
rest
of
the
sound.

No problem . I planned that on the sub somehow .

The signal could be summed via 100k resistors from each anode
of the 5687 and the junction of these R can have a cap to
ground
with
a pole at about 200Hz.
The R junction is also taken to a single CF to make a buffer
before
the
second order filter with pole at say 150, and you will have 18
dB/octave
slope above 300 Hz.

I don't see the need of that 18dB/octave slope , 12dB/octave
should
do
the trick since the satelites will (roughly) have a 12dB/octave
slope
too .
A closed box (Qtc = 0.707) has a 6dB/octave slope and the plan
was
to tune the HP filter (upper part schematic) at the same Fc ,
giving
a
12dB/otave Linkwitz-Riley slope (Q = 0.5) for the satelites .

Frequencies above 100 Hz become more noticeable to the ear
as F rises. The large cone area of a sub makes the F between say
150
and
say 1k
far too noticeable unless you filter them well, hence my use of
a filter with a 24 dB/octave slope by 500 Hz.

The slopes of filters on the satelites don't matter as much and
just to remove their low bass a large series cap may do
for a first order filter.
Satelites make little deep bass, and the bass they do make become
less
noticable
as F falls.
Some overlap is ok.





The only thing missing is a gain control, but the level of
bass
recoverd
as suggested will be higher that fed to your PL86.

Yes , roughly double . No big problem . I planned a TDA2030
chip(s)
for
the sub-amp , bridged and/or boosted with a pair of transistors
......
A soundcard can drive that chip to full output , so I don't need
extra
gain
I
think and so it might be better to take the signal for the mixer
from
the
cathode of the 5687
And if so , what harm can an extra op-amp do if I use the
TDA2030
anyway
.
Maybe a PP EL34 or 6L6 in the future , but I think a TDA2030
will do
just
fine for the rumbling in the sub .....

But for a passive RC type with fairly rapid roll off with
ultimate
attenuation of 12 dB/octave, you can have a low impedance
source
such as a cathode follower driving 10k with 0.068 uF to 0V,
giving a pole at 234Hz.
Then a following RC section with 47k and 0.015 uF for a pole
at
212
Hz.
The second RC filter has 5 times the impedance value of the
first,
so they act to give a reasonably sharper roll off than if you
had
two
RC
filters

with the same values.
The combined -3dB pole will be at around 150 Hz, and
by 500 Hz the slope has become steep.
Its naturally damped, and simple.

Hmm , very odd dB slopes , way to close to Fc , giving way to
odd
phase
shifts I think .

Nope, the passive RC filters have gentle phase shift at the Fc.
A second order filter se to make a flat response followed by as
steep
as possible slope of 12 dB makes 90 degress of phase shift
at -3dB.
And a 24dB /octave similarly "critically damped" filter has 108d
shift at
its -3dBb Fc.




The 12dB/octave Linkwitz-Riley LP filter shown in my schematic
can
simply
be tuned (vary Fc) by turning the dual 100k pot .

What happens to the shape of the response curve?

With passive RC, I use a 3 position switch to change from
say 37, 55 and 90 Hz cut off points by switching the value of C.

No multiway pots are needed.


Yes, but research into sub-woofer powered sports cars with
solar
roof
panels
is underway.

They are training the bass player and drummer to provide the
right
type of sound to use, and design of suitable ear muffs for the
driver
is just about done.
One British horn speaker maker was contracted to provide a
horn
powered
protoype,
but they had trouble getting a permit from the transport
authorities,
since it couldn't fit under the bridges.
Then someone said it'd be more effective to use sub powered
barges,
and
limit
them to use in countryside canals, but the Society Against
Hearing
Damage
in
Tortoises
gained an injunction in the High Court against underwater subs
and underwater horns.

Progress continues in the face of obstacles.

LOL !!!!

Now I understand why Michael Schumachel already has his own
brand of
sub-woofers !!!!

Ah, there's money to be made if you get famous and faster than
everyone
else.
I wish him well and hope he don't do an Ayton Senna Barrier
Interface
Experiment.

Patrick Turner.



Ronald .